ARTICLE IN PRESS Deep-Sea Research I 54 (2007) 1350–1360 www.elsevier.com/locate/dsri The relationship between the standing stock of deep-sea macrobenthos and surface production in the western North Atlantic Nicholas A. Johnsona,Ã, Janet W. Campbellb, Timothy S. Mooreb, Michael A. Rexc, Ron J. Etterc, Craig R. McClaind, Mark D. Dowelle aCommittee on Evolutionary Biology, University of Chicago, Chicago, IL 60637, USA bOcean Process Analysis Laboratory, University of New Hampshire, Durham, NH 03824, USA cDepartment of Biology, University of Massachusetts, Boston, MA 02125, USA dMonterey Bay, Aquarium Research Institute, 7700 Sandholdt Road, Moss Landing, CA 95039, USA eInstitute for Environment and Sustainability, Commission of the European Communities, Joint Research Center, TP 272 I-21020 Ispra, Italy Received 16 August 2006; received in revised form 3 April 2007; accepted 5 April 2007 Available online 29 April 2007 Abstract The relationship between surface production and benthic standing stock is fundamental to understanding biogeography in the deep sea. While much has been learned about the complex oceanographic processes involved in energy transfer to the benthos on local scales, the correspondence of overhead production to benthic community structure on regional scales remains poorly characterized. We compiled a database on the biomass and abundance of deep-sea macrobenthos in the western North Atlantic collected from 1961 to 1985. Using SeaWiFS satellite color imagery, we calculated POC from surface chlorophyll a concentrations (from 1997 to 2001), and estimated POC flux to the seafloor by using the empirically derived Pace et al. [1987. Primary production, new production and vertical flux in the eastern Pacific Ocean. Nature 325, 803–804] algorithm. The standing stock and surface production data are not concurrent, but their basic geographic trends at these very large spatial scales appear to be relatively stable over the time scales of measurement. Estimated POC flux at depth accounts for 62–67% of the variance ðPo0:0001Þ in benthic standing stock, suggesting that macroecological studies of the relationship between satellite-derived surface production and deep-sea community structure may be possible. r 2007 Elsevier Ltd. All rights reserved. Keywords: Deep sea; Macrobenthos; Abundance; Biomass; Surface production; SeaWiFS 1. Introduction ÃCorresponding author. Tel.: +1 773 702 8940; As perhaps the most indirect trophic connection fax: +1 773 702 4699. in nature, the deep-sea benthos is fueled primarily E-mail addresses: [email protected] (N.A. Johnson), by phytodetritus that originates as surface produc- [email protected] (J.W. Campbell), timothy.moore@ unh.edu (T.S. Moore), [email protected] (M.A. Rex), tion and either sinks through the water column or is [email protected] (R.J. Etter), [email protected] laterally advected. The entire process of pelagic– (C.R. McClain), [email protected] (M.D. Dowell). benthic coupling is exceedingly complex and still 0967-0637/$ - see front matter r 2007 Elsevier Ltd. All rights reserved. doi:10.1016/j.dsr.2007.04.011 ARTICLE IN PRESS N.A. Johnson et al. / Deep-Sea Research I 54 (2007) 1350–1360 1351 incompletely understood (Gage, 2003). It is clear POC flux measured at sediment traps moored that carbon flux decreases exponentially with depth 50–600 m above the abyssal seafloor in the northeast (Suess, 1980; Pace et al., 1987; Berelson, 2001); this, Pacific. Sun et al. (2006) found a significant in turn, is assumed to cause the well-known relationship between satellite-derived primary pro- exponential decrease in benthic standing stock with ductivity and Holocene foraminiferan abundance at depth (Rowe, 1983; Heip et al., 2001; Rex et al., lower bathyal and abyssal depths at ocean-wide 2006). This critical, but largely unquantified, scales in the North Atlantic. Here, we examine the assumption is often invoked to explain geographic relationship between surface production and stand- patterns of community structure (reviewed by Levin ing stock of the benthic macrofauna along a depth et al., 2001; Stuart et al., 2003; McClain, 2004; Rex gradient at regional scales in the western North et al., 2005) and evolutionary divergence (Etter and Atlantic. Rex, 1990; Rex and Etter, 1998; McClain et al., 2004; Etter et al., 2005) within the deep sea. At a few 2. Materials and methods sites in the ocean, long-term studies of satellite- derived primary production, sediment-trap mea- 2.1. Benthic standing stock surements of carbon flux, sediment community oxygen consumption, and benthic standing stock We compiled benthic standing stock estimates have documented the major features of vertical from previous studies (Sanders et al., 1965; Rowe energy transfer and its ultimate incorporation by the et al., 1974, 1982; Smith, 1978; Maciolek et al., benthos (e.g., Smith et al., 2006). However, on the 1987a, b) in the western North Atlantic (Fig. 1). larger regional scales at which significant changes in The data represent the macrobenthos, primarily benthic community structure are observed, little is polychaetes, peracarid crustaceans, and molluscs known of the relationship between patterns of retained on 3002420 mm sieves. The benthic sam- overhead production and the abundance of seafloor ples cover depths extending from the shelf-slope life. This relationship is vital to understanding transition down to the abyssal plain (200–5200 m), a biogeography in the deep sea. latitudinal range of 33:1241:3N, and a longitu- It is theoretically possible to estimate the amount dinal range of 65273:8W. Some samples in the of organic material available to the benthos at Rowe et al. (1982) study were collected in Hudson depth. Recent ocean color imagery from the Canyon. Standing stocks appeared to be elevated SeaWiFS satellite provides high-quality chlorophyll only at the canyon head (203–570 m), though no a concentration estimates (Rhea and Davis, 1997) neighboring slope-face samples were available for and makes possible direct computation of surface comparison in that study. Standing stock data POC (Loisel et al., 2002). Sediment trap studies include both abundance (all studies) and wet-weight (Berelson, 2001; Lutz et al., 2002) have generated biomass (all except Maciolek et al., 1987a, b). In empirical predictive models that take surface total, the benthic database includes 127 abundance production as input and describe the rate at which estimates and 80 biomass estimates from samples particulate matter is re-mineralized as it descends collected between 1961 and 1985. The database, through the water column. By combining satellite including sampling locality, type of gear, and sieve color data with these models, estimates of food mesh size used is available as an online supplement. supply to the benthos in terms of POC can be calculated and compared to measurements of 2.2. Surface production benthic standing stock, at least on an annual or longer time scales. POC flux is expected to be a good We acquired global 9 km resolution average predictor of standing stock because it is significantly monthly SeaWiFS satellite data in .hdf format related to benthic community energy demand through the University of New Hampshire Ocean although there are time lags involved (Smith and Process Analysis Laboratory for September 1997 Kaufmann, 1999; Smith et al., 2002). through December 2001. Each pixel represents Despite all of the uncertainty about processes 0:088 latitude by 0:088 longitude, or 9 km  9km involved in pelagic–benthic coupling, Smith et al. at the equator. In the western North Atlantic basin, (2006) recently were able to use satellite-derived each pixel is approximately a 9 km  6 km area. The surface chlorophyll concentrations to calculate data include the following relevant parameters: export flux which was a significant predictor of surface chlorophyll a concentrations ([Chl]), average ARTICLE IN PRESS 1352 N.A. Johnson et al. / Deep-Sea Research I 54 (2007) 1350–1360 Fig. 1. Bathymetric map of the western North Atlantic. Contour lines are every 500 m, with an additional line at 200 m depth to mark the shelf boundary. Station locations for benthic standing-stock data are plotted as white þ symbols. attenuation coefficient for downwelling irradiance events during periods of low productivity (Conte at 490 nm (Kd(490)), and normalized water-leaving et al., 2003). Coastal stations exhibit much more radiance at 490 nm (Lwn(490)). POC was calculated variability over the course of a year than do deep- by using the method of Loisel et al. (2002), though water stations. The nature of the spring bloom is we have used a different relationship between POC discussed at length by Follows and Dutkiewicz and bp (backscattering due to particles) based on (2002). Claustre et al. (1999). An explanation of the calculations is available in the online supplement. 2.3. Scaling The 4-year average of surface [Chl] over the western North Atlantic falls off rapidly with Given the spatial and temporal variations in increasing distance to land (Fig. 2). The pattern of surface production and dispersal of export flux by surface POC, not shown here, is similar. Fig. 2 currents, a fairly large spatial domain must con- demonstrates that most of the highly productive tribute organic carbon to a local benthic site. waters reside over the continental margin. The Deuser et al. (1990) proposed that this domain be western North Atlantic experiences two plankton conceptualized as a statistical funnel. The ocean blooms over the course of the year, in the spring and surface supplying the funnel with export flux is late fall (Campbell and Aarup, 1992; Gregg, 2002). termed its catchment area. Sizes of catchment areas High productivity waters can extend beyond the are still poorly characterized but models of statis- shelf-slope margin during bloom periods (Campbell tical funnels suggest dimensions approaching and Aarup, 1992; Ryan et al., 1999), and as episodic 100 km (Siegel and Deuser, 1997; Siegel and ARTICLE IN PRESS N.A.